In order to focus our attension on the class of applications that this paper concerns itself with, we present an example of a generic distributed multimedia information system; that of multimedia information systems for museums. These multimedia presentations usually share a small number of multimedia servers that they store a mix of continuous and discrete media (such as audio/video and text/images, respectively). One more specific instance is the multimedia information system of a natural history museum where a number of text, images, video and audio data concerning animals have been stored in one storage server (ms1) and a number of text, images, video and audio data concerning plants have been stored in the other storage server (ms2)(see figure 2). Based on this type of local network environment we will focus on two specific Web-based multimedia applications, encoded in the extended-SMIL language and displayed on two different nodes/workstations of the network. Figure 2 contains also the extended-SMIL description of these two multimedia presentations of this natural history museum. The first presentation (L - at the left hand side of the figure) consists of two video streams v1,v2 and one audio stream a1. The second presentation (R - at the right hand side of the figure) started in parallel on a different node of the network by another visitor. It consists of one video stream v3 and one audio stream a2.
Playing the above set of multimedia presentations in a traditional network architecture two main problems are met. Firstly, the best-effort service model provided by the existing systems does not address the temporal dimension of the continuous media data during their retrieval and transmission phase. However, resource reservation even if it is required, it is not the final answer to the end-users. The end-users actually care on how to exploit all the available (and reserved) resources in a best way such that the multimedia application will be presented according to the expected quality requirements. For example, a 10% reservation of the total bandwidth to a video presentation means that the video can be played either colored with a rate of 10 frames per second or grey-scaled with a rate of 18 frames per second. The decision has to be taken by the end-users and the multimedia authors, providing high-level language primitives and special annotation for the definition of any quality requirement.